AI Architect – Generative AI Solutions Lead

About the role

As an AI Architect at PwC you will design and guide production‑ready Generative AI solutions for clients, from discovery through implementation. The role blends hands‑on architecture, technical leadership and close client collaboration.

Key responsibilities

• Lead the architecture and delivery of Generative AI solutions, advising development teams on tools, frameworks and platforms.
• Own scalability, performance, latency, observability, security and cost optimisation across AI solutions.
• Define the appropriate adaptation approach for each use case – prompting, RAG, fine‑tuning or distillation.
• Develop and own evaluation strategies, including golden datasets, LLM‑as‑judge and CI regression testing.
• Set AI security, privacy, Responsible AI and governance controls such as prompt‑injection mitigation and sandboxing.
• Supervise vendor developers, mentor engineers and junior architects, and support business development activities.

Required profile

• 5+ years of experience in agile product delivery and designing production software at scale.
• Demonstrated expertise in Generative AI solution architecture and at least one Generative AI system delivered to production.
• Bachelor’s or Master’s degree in Computer Science, Engineering or equivalent practical experience.
• Preferred Azure certification (e.g., Azure Solutions Architect Expert).

Required skills

• Generative AI models (GPT‑4, Gemini, Llama)
• Retrieval‑augmented generation (RAG) and agent orchestration
• Microsoft Copilots, Azure OpenAI Service, Semantic Kernel, LangChain, AWS Bedrock, Google Vertex AI
• Vector search and retrieval engineering (chunking, hybrid search, reranking)
• Production AI agents (tool calls, state management, durable execution)
• Structured generation with schema‑enforced outputs
• Azure Machine Learning and Azure AI services
• Service‑oriented, event‑driven and microservices architectures
• Microsoft SQL Server, MongoDB
• Docker, Kubernetes
• CI/CD, Azure DevOps, GitHub, Jira
• AI evaluation (golden datasets, LLM‑as‑judge)
• LLMOps observability (LangSmith, Langfuse, Azure Monitor)
• Model routing, caching, batching and cost‑per‑transaction estimation
• AI security controls (prompt injection mitigation, sandboxing, least‑privilege identities)

What we offer

• Professional, positive, team‑oriented environment with international exposure.
• Company training and strong career‑growth opportunities.
• Challenging projects for leading clients.
• Comprehensive benefits: additional medical insurance, food vouchers (€102.26 / month), sport card, fringe benefit and annual bonus.
• Central office in Sofia with the possibility to work from home.